Park et al.

BMC Pregnancy and Childbirth (2018) 18:146

https://doi.org/10.1186/s12884-018-1780-7

RESEARCH ARTICLE Open Access

Plasma inflammatory and immune proteins

as predictors of intra-amniotic infection
and spontaneous preterm delivery
in women with preterm labor:
a retrospective study
Hyunsoo Park1,2, Kyo Hoon Park1,2* , Yu Mi Kim1, Song Yi Kook1,2, Se Jeong Jeon1,2 and Ha-Na Yoo1

Abstract
Background: We investigated whether various inflammatory and immune proteins in plasma predict intra-amniotic
infection and imminent preterm delivery in women with preterm labor and compared their predictive ability with
that of amniotic fluid (AF) interleukin (IL)-6 and serum C-reactive protein (CRP).
Methods: This retrospective cohort study included 173 consecutive women with preterm labor who underwent
amniocentesis for diagnosis of infection and/or inflammation in the AF. The AF was cultured, and assayed for IL-6.
CRP levels and cervical length by transvaginal ultrasound were measured at the time of amniocentesis. The stored
maternal plasma was assayed for IL-6, matrix metalloproteinase (MMP)-9, and complements C3a and C5a using
ELISA kits. The primary and secondary outcome criteria were positive AF cultures and spontaneous preterm delivery
(SPTD) within 48 h, respectively. Univariate, multivariate, and receiver operating characteristic analysis were used for
the statistical analysis.
Results: In bivariate analyses, elevated plasma IL-6 level was significantly associated with intra-amniotic infection and
imminent preterm delivery, whereas elevated plasma levels of MMP-9, C3a, and C5a were not associated with these
two outcomes. On multivariate analyses, an elevated plasma IL-6 level was significantly associated with intra-amniotic
infection and imminent preterm delivery after adjusting for confounders, including high serum CRP levels and short
cervical length. In predicting intra-amniotic infection, the area under the curve (AUC) was significantly lower for plasma
IL-6 than for AF IL-6 but was similar to that for serum CRP. Differences in the AUCs between plasma IL-6, AF IL-6, and
serum CRP were not statistically significant in predicting imminent preterm delivery.
Conclusions: Maternal plasma IL-6 independently predicts intra-amniotic infection in women with preterm labor;
however, it has worse diagnostic performance than that of AF IL-6 and similar performance to that of serum CRP. To
predict imminent preterm delivery, plasma IL-6 had an overall diagnostic performance similar to that of AF IL-6 and
serum CRP. Plasma MMP-9, C3a, and C5a levels could not predict intra-amniotic infection or imminent preterm delivery.
Keywords: Interleukin-6, Intra-amniotic infection, Maternal plasma, Preterm labor, Proteins, Spontaneous preterm delivery

* Correspondence: [email protected]
1
Department of Obstetrics and Gynecology, Seoul National University
College of Medicine, Seoul National University Bundang Hospital, 82, Gumi-ro
173 Beon-gil, Seongnamsi, Kyeonggido 463-707, Korea
2
Center for High-risk Pregnancy and Neonate, Seoul National University
Bundang Hospital, Seongnam, Korea

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Park et al. BMC Pregnancy and Childbirth (2018) 18:146 Page 2 of 9

Background and (4) cervical length measured at the time of amniocen-

Preterm labor and intact membranes are responsible for tesis. The exclusion criteria were: (1) multifetal pregnancy;
approximately one-third of preterm births, and are associ- (2) preterm premature rupture of the membranes
ated with significant perinatal mortality and morbidity [1, (pPROM); (3) prior or subsequent cerclage; (4) active
2]. Adverse neonatal outcomes are closely related to the labor at enrollment (defined as the presence of cervical
occurrence of complications caused by the degree of pre- dilatation greater than 3 cm); (5) evidence of clinical chor-
maturity and intra-uterine infection [3–5]. Therefore, given ioamnionitis; and (6) major fetal congenital anomalies.
a reported intra-amniotic infection rate of 10% [2, 6], a Preterm labor was defined as the presence of regular uter-
more accurate prediction of intra-amniotic infection and ine contractions, with a frequency of at least 2 contrac-
preterm delivery, especially using noninvasive methods, is tions every 10 min, and cervical change (softening,
clinically important for improving pregnancy management effacement, or dilation) that required hospitalization. Ges-
and for counseling women with preterm labor. tational age was determined based on the last menstrual
Traditionally, measuring inflammatory and immune period and the first or second trimester (≤20 weeks) ultra-
proteins in amniotic fluid (AF) is thought to be of value sound results, when available. The ethics committee at
in predicting intra-amniotic infection and preterm birth Seoul National University Bundang Hospital approved the
in women with preterm labor. Several studies have study (IRB no. B-1105/128–102). During the study period,
shown that increased levels of interleukin (IL)-6, matrix amniocentesis for retrieval of AF and maternal blood sam-
metalloproteinase (MMP)-9, and complement split prod- pling were immediately offered to patients who were ad-
ucts (C3a and C4a) in AF are associated with both intra- mitted to our institution. Patients provided written
amniotic infection and preterm parturition [7–11]. How- informed consent for the collection and use of the blood
ever, their clinical use is currently limited by the require- and AF samples and for the use of their clinical informa-
ment of invasive AF sampling. In this regard, tion for research purposes. The primary and secondary
noninvasive approaches that analyze these proteins in a outcome measures were positive AF cultures and SPTD
maternal blood sample may provide a safe alternative to within 48 h, respectively.
traditional invasive amniocentesis because intra-uterine
infection and preterm parturition may lead to a signifi- Amniotic fluid, sample collection and preparation
cant increase in several cytokines simultaneously in both After obtaining informed consent, a transabdominal am-
the AF and maternal blood compartments [12–16]. A niocentesis was performed under sonographic guidance
few studies have reported the role of these inflammatory with a 22-gauge spinal needle under aseptic conditions,
and immune proteins in a maternal blood sample, par- and 15 to 20 mL of AF was aspirated. The AF was imme-
ticularly in preterm labor. Some of these studies are lim- diately sent to the laboratory for culture of aerobic/anaer-
ited by a small sample size [17], a cross-sectional study obic bacteria and genital mycoplasmas (Ureaplasma
design (which could not establish a well-established asso- urealyticum and Mycoplasma hominis) using previously
ciation, unlike a longitudinal design) [18, 19], and no described methods [20]. The remaining AF was centri-
adjustment for confounding factors [13, 19]. This study fuged at 1500 g and 4 °C for 10 min, aliquoted, and stored
aimed to determine whether the various inflammatory at − 70 °C for further analysis. The IL-6 levels in the stored
and immune proteins in plasma are predictive of intra- AF were measured using the enzyme-linked immunosorb-
amniotic infection and imminent preterm delivery (de- ent assay (ELISA) human DuoSet Kit (R&D Systems, Min-
fined as spontaneous preterm delivery [SPTD] within neapolis, MN, USA). The range of the IL-6 standard curve
48 h) in women with preterm labor and to compare their was 7.8–600 pg/mL. The intra- and inter-assay coefficients
predictive ability with that of AF IL-6 and serum C- of variation were < 10% each. CRP and cervical length by
reactive protein (CRP). transvaginal ultrasound were measured in participants at
the time of amniocentesis as part of the hospital protocol
Methods using methods that have been previously described [20,
Study population 21]. The results of AF culture, CRP, and cervical length
A retrospective cohort study was conducted on consecu- measurement were available to the managing clinicians,
tive singleton pregnant women diagnosed with preterm but the AF IL-6 results were not.
labor (23 + 0 to 33 + 6 weeks of gestation) who were
admitted to Seoul National University Bundang Hospital Inflammatory and immune mediator assays in plasma
(Seongnamsi, Republic of Korea) from June 2004 to At the time of amniocentesis, maternal blood sam-
April 2015. The inclusion criteria were: (1) a live fetus; ples were obtained from all participants and were
(2) an aliquot of maternal plasma available for analysis; collected in EDTA tubes. The samples were centri-
(3) amniocentesis performed to evaluate the infectious fuged at 1500 g at 4 °C for 10 min, and the super-
and inflammatory status of AF at the time of enrollment; natant was aliquoted and stored at − 70 °C until
Park et al. BMC Pregnancy and Childbirth (2018) 18:146 Page 3 of 9

assayed. The stored plasma samples were assayed for after adjusting for baseline variables, with a P value < 0.1
multiple inflammatory and immune proteins (IL-6, MMP- in univariate analysis. Independent variables associated
9, C3a and C5a). ELISA kits were used to measure IL-6 with invasive amniocentesis were not included in the
(R&D System, Minneapolis, MN, USA), MMP-9 (DuoSet logistic regression model, because these were not
ELISA from R&D System, Minneapolis, MN, USA) and baseline variables but were used for comparison. In the
complement C3a and C5a (BD Biosciences, San Diego, CA logistic regression model, continuous data were
92121, USA) levels in the plasma samples, according to the transformed into dichotomous data to decrease the
manufacturer’s instructions. The ranges of the IL-6, MMP- problem of multicollinearity, especially between the
9, C3a, and C5a standard curves were 0.2–10 pg/mL, 31.2– serum CRP and plasma IL-6 (r = 0.495) or for prediction
2000 pg/mL, 0.078–2.5 ng/mL, and 0.08–2.5 ng/mL, or decision-making purposes. Receiver-operating charac-
respectively. The intra- and inter-assay coefficients of teristic (ROC) curves were used to identify the best cut-
variation were 3.6 and 10.9% for IL-6, 2.9 and 18.3% for off values for dichotomization. The optimal cut-off
MMP-9, 10.4 and 13.7% for C3a, and 7.9 and 8.0% for C5a, values were selected on the basis of the Youden index
respectively. (maximum [sensitivity + specificity] − 1). The areas
under the ROC curves (AUCs) for each protein and the
Management of preterm labor and definitions of various cervical length were measured, and compared using the
factors method described by DeLong et al. [24]. The sensitivity,
Intra-amniotic infection was defined as the presence of a specificity, and predictive values of plasma IL-6 and cer-
positive AF culture for microorganisms. The patients vical length with reference to AF IL-6 and serum CRP
were hydrated and, if uterine contractions persisted, they levels were calculated and compared using the McNe-
were started on a regimen of intravenous tocolysis with mar test. The sampling-to-delivery interval was assessed
magnesium sulfate, ritodrine, or atosiban. Corticoste- using Kaplan–Meier analysis and was compared between
roids were administered between 24 and 34 weeks of groups using the log-rank test. The Cox proportional
gestation. The decision to use prophylactic antibiotics model was used to determine whether higher levels of
was left to the discretion of the attending obstetrician. various proteins in plasma were associated with the
At our institution, antibiotic treatment of women with sampling-to-delivery interval after adjusting for other
preterm labor for pregnancy prolongation is not recom- prognostic variables. The Spearman rank correlation test
mended except for group B streptococcus prophylaxis or was used to measure the relationship between the
treatment of cases with clinically suspected (AF WBC continuous variables that did not follow a normal distri-
count ≥ 50 cells/mm3) or diagnosed intra-amniotic infec- bution. All P values were two-sided, and P < .05 was
tion, as previously described in detail [20]. Although the considered statistically significant.
regimen for antibiotic prophylaxis in these patients var-
ied according to the study period and the obstetrician, Results
ampicillin and azithromycin (clarithromycin or erythro- In total, 173 women with a diagnosis of preterm labor
mycin) were the main antibiotics used. For women with who met all the inclusion criteria were finally included.
intra-amniotic infection, antibiotics are selected on the Their mean gestational ages (± standard deviation) at
basis of the type of microorganisms found in the AF cul- sampling and at delivery were 30.1 ± 2.7 weeks and 35.3
tures. Medications such as antibiotics, corticosteroids, ± 3.9 weeks, respectively. The prevalence of intra-
and tocolytics were started after sampling. A histologic amniotic infection was 9.2% (16/173), and the microbes
diagnosis of chorioamnionitis was made in accordance isolated from the AF in 16 patients were Ureaplasma
with the definition previously described in detail [22]. urealyticum (n = 16) and Mycoplasma hominis (n = 13).
Clinical chorioamnionitis was diagnosed in accordance Polymicrobial invasion was present in 13 of the 16 pa-
with the criteria proposed by Gibbs et al. [23]. tients (81.2%). SPTD within 48 h of sampling occurred
in 15.0% (26/173) of the patients.
Statistical methods Of the measured plasma proteins (IL-6, MMP-9, C3a,
Data were analyzed using SPSS version 22.0 for Win- and C5a), positive significant correlations were found
dows (IBM Corp., Armonk, NY, USA). The Shapiro– only between C5a and C3a (r = 0.376, P < 0.001) and
Wilk test was used to determine whether the data in between C5a and MMP-9 (r = 0.166, P = 0.029), and only
groups were distributed normally. Univariate analysis the plasma IL-6 level was significantly correlated with
was performed using the Student’s t-test or Mann–Whit- serum CRP level (r = 0.494, P < 0.001) and AF IL-6 level
ney U test for continuous data and the χ2-test or Fisher’s (r = 0.289, P < 0.001). Cervical length was significantly
exact test for dichotomous data. A multivariate logistic correlated with the plasma MMP-9 level (r = − 0.269,
regression model was used to examine the relationship P < 0.001) and serum CRP (r = − 0.159, P = 0.036), while
of the level of each protein in plasma to the outcome other immune-related proteins measured in plasma
Park et al. BMC Pregnancy and Childbirth (2018) 18:146 Page 4 of 9

were not correlated with cervical length (all variables, age (≥33 years); multiparity, high level of serum CRP
r = − 0.114 to − 0.040, P > 0.05). (≥3.5 mg/L), high level of plasma IL-6 (≥4.8 pg/mL), and
The demographic and clinical characteristics of the short cervical length (≤20.0 mm). Logistic regression
study population according to AF culture results are pre- showed that a high plasma IL-6 level was significantly
sented in Table 1. Table 2 describes the clinical charac- associated with imminent preterm delivery, even after
teristics of the study population according to the adjustment for other confounders, including high serum
occurrence of SPTD within 48 h after sampling. CRP levels, short cervical length, and multiparity (Table 3).
Multiple logistic regression analyses were performed Fig. 1 displays the ROC curves for the plasma IL-6, AF
to further examine the relationship between the various IL-6, and serum CRP levels in predicting intra-amniotic
proteins in plasma, intra-amniotic infection, and infection (Fig. 1a) and imminent preterm delivery (Fig. 1b).
imminent preterm delivery after adjusting for the effects In predicting intra-amniotic infection, the AUC for plasma
of baseline variables. Regarding the prediction of intra- IL-6 was significantly lower than that for AF IL-6
amniotic infection, the following variables were entered (P < 0.001), but was not different from the AUC for serum
into the multivariate logistic regression analysis as sig- CRP (P = 0.414) (Fig. 1a). Differences in the AUCs between
nificant predictors in the univariate analyses (P < 0.1): plasma IL-6, AF IL-6, and serum CRP tests were not
gestational age at sampling, serum CRP and plasma IL-6 statistically significant in predicting imminent preterm
levels, and cervical length. In this model, all continuous delivery (Fig. 1b).
predictors were entered as dichotomous variables using Figure 2 shows the Kaplan–Meier estimates of the
the cut-off values derived from the ROC curves. The sampling-to-delivery interval for a plasma IL-6 level of
optimal cutoff values for gestational age at sampling, ≥4.8 or < 4.8 pg/mL (Fig. 2a), serum CRP level of ≥3.5 or
serum CRP level, plasma IL-6 level, and cervical < 3.5 mg/L (Fig. 2b), and AF IL-6 level of ≥1.5 or < 1.5 ng/
length were ≤ 29.0 weeks, ≥11.2 mg/L, ≥4.8 pg/mL, mL (Fig. 2c). Comparisons made by the log rank test were
and ≤ 20.6 mm, respectively. A multivariate logistic regres- statistically significant (plasma IL-6 ≥ 4.8 pg/mL, P < 0.001;
sion model indicated that only a high plasma IL-6 level was CRP ≥3.5 mg/L, P = 0.001; and AF IL-6 ≥ 1.5 ng/mL,
independently associated with intra-amniotic infection P < 0.001). Similarly, using Cox proportional hazards
(Table 3). Likewise, when imminent preterm delivery was regression model, high level of plasma IL-6 was found to
used as an outcome measure, the following dichotomized be significantly associated with the sampling-to-delivery
variables were used for logistic regression analysis: older interval after being controlled for multiparity, high level of

Table 1 Characteristics of the study population according to amniotic fluid culture results
Amniotic fluid culture P-value
Positive (n = 16) Negative (n = 157)
Maternal age (years) 31.3 ± 4.9 31.6 ± 4.0 0.638
Nulliparity 56% (9) 62% (97) 0.666
Gestational age at sampling (weeks) 28.3 ± 3.0 30.2 ± 2.6 0.014
Gestational age at delivery (weeks) 30.0 ± 3.1 35.9 ± 3.5 < 0.001
Serum C-reactive protein (mg/L) 25.3 ± 21.5 7.5 ± 11.6 < 0.001
Cervical length on ultrasonography (mm) 19.2 ± 8.9 25.3 ± 12.1 0.054
Plasma C3a (μg/mL) 13.8 ± 6.8 12.8 ± 7.9 0.366
Plasma C5a (ng/mL) 46.0 ± 10.3 44.8 ± 13.0 0.712
Plasma interleukin-6 (pg/mL) 9.9 ± 9.3 4.6 ± 6.1 0.001
Plasma matrix metalloproteinase-9 (ng/ml) 248.7 ± 232.0 210.6 ± 185.0 0.405
Use of tocolytic agents 94% (15) 94% (147) 0.985
Use of antibiotics 75% (12) 26% (40) < 0.001
Use of antenatal corticosteroids 100% (17) 75% (120) 0.026
Clinical chorioamnionitis 6% (1) 0% (0) 0.002
Histological chorioamnionitisa 75% (12/16) 32% (27/85) 0.001
Amniotic fluid interleukin-6 (ng/mL) 26.3 ± 21.4 1.9 ± 7.7 < 0.001
Data are given as mean ± standard deviation or % (n/N)
a
Data for the histologic evaluation of the placenta were only available for 101 of the 173 women because delivery took place at another institution in 5 cases, and
histologic evaluation of the placenta was not performed in 67 cases because of our institutional policy that only the placentas from preterm deliveries are sent for
histopathologic examination
Park et al. BMC Pregnancy and Childbirth (2018) 18:146 Page 5 of 9

Table 2 Characteristics of the study population according to the occurrence of delivery within 48 h
Spontaneous preterm delivery after sampling P-value
≤48 h (n = 26) > 48 h (n = 147)
Maternal age (years) 33.0 ± 4.0 31.3 ± 4.0 0.047
Nulliparity 35% (9) 66% (97) 0.003
Gestational age at sampling (weeks) 30.7 ± 2.6 30.0 ± 2.7 0.195
Gestational age at delivery (weeks) 30.7 ± 2.5 36.1 ± 4.0 < 0.001
Serum C-reactive protein (mg/L) 14.3 ± 15.2 8.2 ± 13.4 0.003
Cervical length on ultrasonography (mm) 13.8 ± 10.9 26.7 ± 11.1 < 0.001
Plasma C3a (μg/mL) 12.9 ± 7.9 12.9 ± 7.7 0.863
Plasma C5a (ng/mL) 41.0 ± 14.0 45.6 ± 12.5 0.237
Plasma interleukin-6 (pg/mL) 11.7 ± 12.0 4.0 ± 4.2 < 0.001
Plasma matrix metalloproteinase-9 (ng/mL) 251.0 ± 253.9 207.6 ± 175.9 0.858
Use of tocolytic agents 89% (23) 95% (139) 0.242
Use of antibiotics 27% (7) 31% (45) 0.706
Use of antenatal corticosteroids 89% (23) 76% (112) 0.165
Clinical chorioamnionitis 3.8% (1) 0% (0) 0.017
Histological chorioamnionitisa 50% (13/26) 35% (26/75) 0.169
Positive amniotic fluid cultures 19% (5) 7.5% (11) 0.057
Amniotic fluid interleukin-6 (ng/mL) 11.7 ± 21.4 2.8 ± 8.9 < 0.001
Data are given as mean ± standard deviation or % (n/N)
a
Data for the histologic evaluation of the placenta were only available for 101 of the 173 women because delivery took place at another institution in 5 cases, and
histologic evaluation of the placenta was not performed in 67 cases because of our institutional policy that only the placentas from preterm deliveries are sent for
histopathologic examination

serum CRP, short cervical length, and antenatal corti- (P = 0.005), and equivalent in sensitivity to CRP
costeroids (hazard ratio: 2.04; 95% confidence interval, (P = 0.687) for detecting intra-amniotic infection. For
1.36–3.05, P = 0.001). detection of imminent preterm delivery, the sensitivity
Table 4 shows the diagnostic values of plasma IL-6, AF of plasma IL-6 was not significantly different from that
IL-6, serum CRP, and cervical length for predicting of AF IL-6 (P = 0.508) or CRP (P = 0.289), whereas the
intra-amniotic infection and imminent preterm delivery. specificity of plasma IL-6 was significantly better than
Plasma IL-6 was less sensitive and specific than AF IL-6 that of CRP (P < 0.001) but not significantly different
(McNemar’s test, P < 0.001), less specific than CRP from that of AF IL-6 (P = 1.000).

Table 3 Multivariate logistic regression of potential predictors of intra-amniotic infection and imminent spontaneous preterm
deliverya
Predictors Odds ratios 95% Confidence interval P-value
Intra-amniotic infection
Early gestational age at sampling (≤ 29.0 weeks) 3.212 0.909–11.354 0.070
Very high serum CRP level (≥ 11.2 mg/L) 3.459 0.937–12.770 0.063
High plasma interleukin-6 level (≥ 4.8 pg/mL) 4.573 1.184–17.661 0.027
Short cervical length (≤ 20.6 mm) 2.492 0.733–8.472 0.144
Imminent preterm delivery
Older age (≥ 33 years) 1.841 0.593–5.716 0.291
Multiparity 3.888 1.242–12.175 0.020
High serum CRP level (≥ 3.5 mg/L) 5.275 1.145–24.299 0.033
High plasma interleukin-6 level (≥ 4.8 pg/mL) 6.515 2.114–20.078 0.001
Short cervical length (≤ 20.0 mm) 8.334 2.591–26.812 < 0.001
CRP C-reactive protein
a
All continuous predictors were entered as dichotomous variables using the cut-off values derived from the receiver-operating characteristic curves to predict
each outcome
Park et al. BMC Pregnancy and Childbirth (2018) 18:146 Page 6 of 9

Fig. 1 Receiver-operating characteristic curves in predicting (a) intra-amniotic infection and (b) imminent preterm delivery. [(a): P < 0.001 between
plasma IL-6 and AF IL-6; P = 0.414 between plasma IL-6 and CRP; (b): no differences (P > 0.1) between plasma IL-6, AF IL-6, and CRP]. IL, interleukin;
AF, amniotic fluid; CRP, C-reactive protein

Discussion women with preterm labor elicits a much weaker inflam-

Our principal findings are as follows: (i) in women with matory response (mediated by IL-6) in the maternal
preterm labor, plasma IL-6 is moderately predictive of plasma compartment than in the AF compartment and,
intra-amniotic infection and imminent preterm delivery; thus, the inflammatory biomarkers in the maternal
(ii) plasma IL-6 has a worse overall diagnostic perform- plasma cannot precisely reflect the infectious status of
ance than that of AF IL-6 and a similar diagnostic per- the amniotic cavity. In fact, these observations are
formance to that of CRP (a prototype of inflammatory natural because microorganisms in the amniotic cavity
response) in predicting intra-amniotic infection; (iii) for can directly increase the production of the chemical
prediction of imminent preterm delivery, the diagnostic mediators in the original site (i.e., AF), where the host
performance of plasma IL-6, AF IL-6, and serum CRP immune response to microorganisms occurs. Accord-
was similar; and (iv) the plasma levels of MMP-9, C3a, ingly, to determine maternal plasma biomarkers, future
and C5a could not predict intra-amniotic infection or studies based on histologic chorioamnionitis rather than
imminent preterm delivery. Similar results regarding the MIAC are warranted because acute inflammatory
relationship between plasma pro-inflammatory cytokines processes of the placenta are of maternal origin [26] and
levels, intra-amniotic infection, and histologic chorioam- may be reflected in the plasma.
nionitis were also documented in women with preterm As previously reported [17, 19, 27, 28], we found that
labor [13, 19, 25] and pPROM [13, 16]. elevated plasma IL-6 levels predict imminent preterm
We found that an elevated plasma level of IL-6 pre- delivery with a diagnostic performance similar to that of AF
dicted intra-amniotic infection in women with preterm IL-6. Further, multivariate logistic regression analysis
labor. Our findings confirm those previously reported by indicated that factors such as multiparity, high serum CRP
Cobo et al., demonstrating that women with preterm level, and short cervical length, as well as high plasma IL-6
labor (22 + 0 to 31 + 6 weeks of gestation) and intra- levels, were also independently associated with an increased
amniotic infection had higher levels of maternal serum risk of imminent preterm delivery, similar to the findings of
IL-6 than did those without intra-amniotic infection Tsiartas et al. [29]. Taken together, these findings suggested
[13]. Moreover, we showed that the association between that the etiology of preterm labor leading to SPTD may be
plasma IL-6 and intra-amniotic infection remained un- multifactorial. Thus, a combination of factors contributing
changed in multivariate logistic analysis after adjusting to the pathophysiology of preterm birth may enhance the
for gestational age at sampling and CRP. Moreover, to ability to predict SPTD. We believe that, unlike predicting
the best of our knowledge, our study is the first to com- intra-amniotic infection, a highly sensitive and specific test
pare plasma IL-6 with AF IL-6 and serum CRP in terms for SPTD can be created by developing novel biomarkers
of overall accuracy in identifying intra-amniotic infec- in maternal plasma, alone or in combination with clinical
tion. We found that plasma IL-6 has a worse overall parameters, thus allowing targeted interventions.
diagnostic performance than that of AF IL-6 in predict- Importantly, we found that serum CRP (the most used
ing intra-amniotic infection. These findings suggest that infection/inflammation marker) has a predictive power
microbial invasion of the amniotic cavity (MIAC) in similar to that of plasma IL-6 for predicting intra-
Park et al. BMC Pregnancy and Childbirth (2018) 18:146 Page 7 of 9

amniotic infection and is similar to that of plasma IL-6

and AF IL-6 for predicting imminent preterm delivery.
Similarly, our previous study found that serum CRP has
a diagnostic performance similar to that of IL-6 and
MMP-9 in AF for the detection of histologic chorioam-
nionitis [7]. These may be important from a clinical
perspective, as the measurement of serum CRP is
simple, noninvasive, and inexpensive and may provide
information to improve maternal and infant health
and patient counseling. However, it is not a sensitive
or specific marker for either intra-amniotic infection
or SPTD [30, 31].
MMP-9 is an important zinc-dependent protease in
the MMP superfamily, which is involved in inflamma-
tion, cytokine processing, and tissue remodeling [32].
Therefore, several studies have investigated the high ex-
pression of MMP-9 in AF associated with intra-amniotic
infection, and found AF MMP-9 levels to be the stron-
gest predictor of intra-amniotic infection in preterm
labor [7, 8]. However, to our knowledge, no studies have
investigated the changes in MMP-9 levels in maternal
plasma in relation to intra-amniotic infection and
imminent preterm delivery in the setting of preterm
labor. We observed that intra-amniotic infection was
not associated with a change in plasma MMP-9 levels.
Moreover, regarding imminent preterm delivery, no
significant change in MMP-9 levels was observed,
consistent with the findings of a previous study
conducted in an asymptomatic cohort [27].
The complement system is pivotal for innate
immunity, the first line of defense against invading path-
ogens [33]. Thus, the relationship between complement
activation, preterm birth, and intra-uterine infection/in-
flammation has been investigated. Interestingly, Lynch
et al. found that elevated plasma levels of the comple-
ment activation fragment C3a and Bb in early gestation
(< 20 weeks) are independent predictive factors for SPTD
[34, 35]. Soto et al. reported that women with preterm
labor and intra-amniotic infection have higher median
maternal plasma C3a and C5a levels than those with
preterm labor without intra-amniotic infection, whereas
there was no difference in the plasma C3a and C5a
levels between women with preterm labor who delivered
at term and those with preterm delivery (< 37 weeks) [18].
However, our study did not reveal any association be-
tween plasma complement C3a/C5a and intra-amniotic
infection or imminent preterm delivery, suggesting that
Fig. 2 Kaplan-Meier survival estimates of the sampling-to-delivery
the maternal immune system in the plasma compart-
interval for (a) plasma IL-6 of ≥4.8 or < 4.8 pg/mL (median, 5.00 days
[95% CI, 0.70–9.30] vs. 45.00 days [95% CI, 35.31–54.69]; P < 0.001), (b) ment related to complement does not respond to the
serum CRP of ≥3.5 or < 3.5 mg/L (median, 19.00 days [95% CI, microbes or microbial products in the amniotic cavity
5.39–32.62] vs. 50.00 days [95% CI, 39.83–60.17]; P = 0.001), and (c) and is not implicated in the process of preterm partur-
AF IL-6 of ≥1.5 or < 1.5 ng/mL (median, 5.00 days [95% CI, 1.99–8.01] ition. This discrepancy may be attributed to differences
vs. 47.00 days [95% CI, 38.86–55.14]; P < 0.001). IL, interleukin; CI,
in the populations studied (asymptomatic women ver-
confidence interval; CRP, C-reactive protein; AF, amniotic fluid
sus those presenting with symptoms of preterm labor),
Park et al. BMC Pregnancy and Childbirth (2018) 18:146 Page 8 of 9

Table 4 Diagnostic indices of various factors to predict intra-amniotic infection and imminent preterm delivery
Variables Area (±SE) under the ROC curve 95% CI Cut-off valuea Sensitivityb Specificityb PPV NPV
(95% CI) (95% CI)
Intra-amniotic infection
Plasma IL-6 (pg/mL) 0.76 ± 0.07 0.63–0.89 4.8 75.0 (50.1–93.2) 75.6 (47.6–92.7) 24.0 96.8
AF IL-6 (ng/mL) 0.98 ± 0.01c 0.96–1.00 3.6 100 (79.4–100.0)d 92.4 (87.0–96.0)d 57.1 100
Serum CRP (mg/L) 0.81 ± 0.06 0.69–0.93 11.2 62.5 (35.4–84.8) 86.0 (80.0–91.0)e 31.3 95.7
Cervical length (mm) 0.66 ± 0.06 0.55–0.77 20.6 62.5 (35.4–84.8) 65.0 (57.0–72.4)f 15.4 94.4
Imminent preterm delivery
Plasma IL-6 (pg/mL) 0.77 ± 0.06 0.66–0.88 4.8 73.1 (52.2–88.4) 78.2 (70.7–84.6) 37.3 94.3
AF IL-6 (ng/mL) 0.74 ± 0.05 0.64–0.85 1.5 61.5 (40.6–79.8) 78.2 (70.7–84.6) 33.3 92.0
Serum CRP (mg/L) 0.68 ± 0.06 0.57–0.79 3.5 88.5 (69.9–97.6) 51.0 (42.7–59.4)d 24.2 96.2
Cervical length (mm) 0.81 ± 0.05 0.72–0.90 20.0 76.9 (56.4–91.0) 71.4 (63.4–78.6) 32.3 94.6
SE standard error; ROC receiver-operating characteristic, CI confidence interval, PPV positive predictive value, NPV negative predictive value; IL-6 interleukin-6,
AF amniotic fluid, CRP C-reactive protein
a
Cut-off values corresponding to the highest sum of sensitivity and specificity
b
Values are given as % (95% CI)
c
P < 0.001 compared to plasma IL-6 by the method of DeLong et al.
d
P < 0.001 compared to plasma IL-6 by McNemar’s test
e
P < 0.01 compared to plasma IL-6 by McNemar’s test
f
P < 0.05 compared to plasma IL-6 by McNemar’s test

study design (cross-sectional versus longitudinal), and C5a levels did not predict intra-amniotic infection or
sample size. imminent preterm delivery. Therefore, the assessment of
Our study has several limitations. First, it was a retro- inflammatory and immune markers in the maternal
spective study conducted at a single center, limiting the plasma may have limited clinical usefulness for identify-
generalizability of our findings. Second, we did not fully ing intra-amniotic infection in women with preterm
characterize the inflammatory and immune proteins in labor. Further studies should utilize cervicovaginal
plasma but only measured the relevant proteins that had fluid, rather than maternal plasma, to help identify
the most potential as a plasma biomarker of high accur- new non-invasive biomarkers for intra-amniotic
acy to detect intra-amniotic infection and preterm birth infection.
[13, 17, 18, 29, 34]. Third, in line with our previous stud-
ies [6, 20], only Ureaplasma urealyticum and Myco- Abbreviations
plasma hominis were detected in the AF, although the AF: Amniotic Fluid; AUC: Area Under the Curve; CI: Confidence Interval;
reason is unclear. However, considering the fact that CRP: C-Reactive Protein; ELISA: Enzyme-Linked Immunosorbent Assay;
IL: Interleukin; MIAC: Microbial Invasion of the Amniotic cavity; MMP: Matrix
various aerobic and anaerobic bacteria were recovered Metalloproteinase; NPV: Negative Predictive Value; pPROM: Premature
from the AF of women with pPROM in our hospital Rupture of the Membranes; PPV: Positive Predictive Value; ROC: Receiver-
during nearly the same period [36, 37], we believe that Operating Characteristic; SE: Standard Error; SPTD: Spontaneous Preterm
Delivery
isolation of only mycoplasma in the current study is not
associated with the technical problems of aerobic and
Funding
anaerobic cultures in our hospital. Fourth, the treatment This study was supported by a grant from the Korea Health Technology R&D
regimen for management of preterm labor varied Project, Ministry of Health and Welfare, Republic of Korea (Grant No. HI
according to the study period and the obstetrician, which 14C1798).
may lead to affecting the outcome of interest and
statistical analyses. Availability of data and materials
The dataset for this study is available from the corresponding author upon
reasonable request.
Conclusions
In conclusion, maternal plasma IL-6 independently
predicts intra-amniotic infection in women with preterm Authors’ contributions
HP: Data acquisition; analysis and interpretation; drafting the article. KHP:
labor; however, it has worse diagnostic performance than Conception and design; data acquisition; analysis and interpretation; drafting
that of AF IL-6 and similar performance to that of the article and final approval of the version to be published. YMK: Data
serum CRP. For prediction of imminent preterm deliv- acquisition and analysis. SYK: Data acquisition; analysis and interpretation. SJJ:
Data acquisition; analysis and interpretation. HNY: Conception and design;
ery, the diagnostic performance of plasma IL-6, AF IL-6, data acquisition; analysis and interpretation. All authors approved the final
and serum CRP were similar. Plasma MMP-9, C3a, and version of the manuscript.
Park et al. BMC Pregnancy and Childbirth (2018) 18:146 Page 9 of 9

Ethics approval and consent to participate 18. Soto E, Romero R, Richani K, Espinoza J, Nien JK, Chaiworapongsa T, et al.
The ethics committee at Seoul National University Bundang Hospital Anaphylatoxins in preterm and term labor. J Perinat Med. 2005;33:306–13.
approved the study (IRB no. B-1105/128–102). Patients provided written 19. Greig PC, Murtha AP, Jimmerson CJ, Herbert WN, Roitman-Johnson B, Allen
informed consent for the collection and use of the blood and AF samples J. Maternal serum interleukin-6 during pregnancy and during term and
and for the use of their clinical information for research purposes. preterm labor. Obstet Gynecol. 1997;90:465–9.
20. Lee SY, Park KH, Jeong EH, Oh KJ, Ryu A, Kim A. Intra-amniotic infection/
Competing interests inflammation as a risk factor for subsequent ruptured membranes after
The authors declare that they have no competing interests. clinically indicated amniocentesis in preterm labor. J Korean Med Sci. 2013;
28:1226–32.
21. Suh YH, Park KH, Hong JS, Noh JH. Prediction of prolonged pregnancy
Publisher’s Note in nulliparous women by transvaginal ultrasonographic measurement of
Springer Nature remains neutral with regard to jurisdictional claims in cervical length at 20-24 weeks and 37 weeks. J Korean Med Sci. 2007;
published maps and institutional affiliations. 22:89–93.
22. Yoon BH, Romero R, Kim CJ, Jun JK, Gomez R, Choi JH, et al. Amniotic fluid
Received: 28 September 2017 Accepted: 26 April 2018 interleukin-6: a sensitive test for antenatal diagnosis of acute inflammatory
lesions of preterm placenta and prediction of perinatal morbidity. Am J
Obstet Gynecol. 1995;172:960–70.
References 23. Gibbs RS, Blanco JD, St Clair PJ, Castaneda YS. Quantitative bacteriology of
1. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of amniotic fluid from women with clinical intraamniotic infection at term.
preterm birth. Lancet. 2008;371:75–84. J Infect Dis. 1982;145:1–8.
2. Romero R, Chaiworapongsa T, Espinoza J. Micronutrients and intrauterine 24. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under
infection, preterm birth and the fetal inflammatory response syndrome. two or more correlated receiver operating characteristic curves: a
J Nutr. 2003;133(Suppl 2):1668S–73S. nonparametric approach. Biometrics. 1988;44:837–45.
3. Bulletins-Obstetrics ACoP. ACOG practice bulletin no. 80: premature rupture 25. Boggess KA, Greig PC, Murtha AP, Jimmerson CE, Herbert WN. Maternal
of membranes. Clinical management guidelines for obstetrician- serum granulocyte-colony stimulating factor in preterm birth with
gynecologists. Obstet Gynecol. 2007;109:1007–19. subclinical chorioamnionitis. Am J Reprod Immunol. 1997;33:45–52.
4. Jung EY, Park KH, Han BR, Cho SH, Yoo HN, Lee J. Amniotic fluid infection, 26. McNamara MF, Wallis T, Qureshi F, Jacques SM, Gonik B. Determining
cytokine levels, and mortality and adverse pulmonary, intestinal, and the maternal and fetal cellular immunologic contributions in preterm
neurologic outcomes in infants at 32 Weeks’ gestation or less. J Korean Med deliveries with clinical or subclinical chorioamnionitis. Infect Dis Obstet
Sci. 2017;32:480–7. Gynecol. 1997;5:273–9.
5. Combs CA, Gravett M, Garite TJ, Hickok DE, Lapidus J, Porreco R, et al. 27. Sorokin Y, Romero R, Mele L, Wapner RJ, Iams JD, Dudley DJ, et al. Maternal
Amniotic fluid infection, inflammation, and colonization in preterm labor serum interleukin-6, C-reactive protein, and matrix metalloproteinase-9
with intact membranes. Am J Obstet Gynecol. 2014;210:125 e1–e15. concentrations as risk factors for preterm birth <32 weeks and adverse
6. Park JW, Park KH, Lee SY. Noninvasive prediction of intra-amniotic infection neonatal outcomes. Am J Perinatol. 2010;27:631–40.
and/or inflammation in women with preterm labor: various cytokines in 28. Maeda K, Matsuzaki N, Fuke S, Mitsuda N, Shimoya K, Nakayama M, et al.
cervicovaginal fluid. Reprod Sci. 2013;20:262–8. Value of the maternal interleukin 6 level for determination of histologic
7. Oh KJ, Park KH, Kim SN, Jeong EH, Lee SY, Yoon HY. Predictive value of chorioamnionitis in preterm delivery. Gynecol Obstet Investig. 1997;43:225–31.
intra-amniotic and serum markers for inflammatory lesions of preterm 29. Tsiartas P, Holst RM, Wennerholm UB, Hagberg H, Hougaard DM,
placenta. Placenta. 2011;32:732–6. Skogstrand K, et al. Prediction of spontaneous preterm delivery in women
8. Harirah H, Donia SE, Hsu CD. Amniotic fluid matrix metalloproteinase-9 with threatened preterm labour: a prospective cohort study of multiple
and interleukin-6 in predicting intra-amniotic infection. Obstet Gynecol. proteins in maternal serum. BJOG. 2012;119:866–73.
2002;99:80–4. 30. Wei SQ, Fraser W, Luo ZC. Inflammatory cytokines and spontaneous
9. Soto E, Romero R, Richani K, Yoon BH, Chaiworapongsa T, Vaisbuch E, et al. preterm birth in asymptomatic women: a systematic review. Obstet
Evidence for complement activation in the amniotic fluid of women with Gynecol. 2010;116:393–401.
spontaneous preterm labor and intra-amniotic infection. J Matern Fetal 31. Yoon BH, Yang SH, Jun JK, Park KH, Kim CJ, Romero R. Maternal blood C-
Neonatal Med. 2009;22:983–92. reactive protein, white blood cell count, and temperature in preterm labor:
10. Elimian A, Figueroa R, Canterino J, Verma U, Aguero-Rosenfeld M, Tejani N. a comparison with amniotic fluid white blood cell count. Obstet Gynecol.
Amniotic fluid complement C3 as a marker of intra-amniotic infection. 1996;87:231–7.
Obstet Gynecol. 1998;92:72–6. 32. Stamenkovic I. Extracellular matrix remodelling: the role of matrix
11. Vogel I, Thorsen P, Curry A, Sandager P, Uldbjerg N. Biomarkers for the metalloproteinases. J Pathol. 2003;200:448–64.
prediction of preterm delivery. Acta Obstet Gynecol Scand. 2005;84:516–25. 33. Walport MJ. Complement. First of two parts. N Engl J Med. 2001;344:
12. Dulay AT, Buhimschi IA, Zhao G, Bahtiyar MO, Thung SF, Cackovic M, et al. 1058–66.
Compartmentalization of acute phase reactants Interleukin-6, C-reactive 34. Lynch AM, Gibbs RS, Murphy JR, Giclas PC, Salmon JE, Holers VM. Early
protein and Procalcitonin as biomarkers of intra-amniotic infection and elevations of the complement activation fragment C3a and adverse
chorioamnionitis. Cytokine. 2015;76:236–43. pregnancy outcomes. Obstet Gynecol. 2011;117:75–83.
13. Cobo T, Tsiartas P, Kacerovsky M, Holst RM, Hougaard DM, Skogstrand K, 35. Lynch AM, Gibbs RS, Murphy JR, Byers T, Neville MC, Giclas PC, et al.
et al. Maternal inflammatory response to microbial invasion of the amniotic Complement activation fragment Bb in early pregnancy and spontaneous
cavity: analyses of multiple proteins in the maternal serum. Acta Obstet preterm birth. Am J Obstet Gynecol. 2008;199:354 e1–8.
Gynecol Scand. 2013;92:61–8. 36. Jung EY, Park KH, Han BR, Cho SH, Ryu A. Measurement of interleukin 8
14. Chow SS, Craig ME, Jones CA, Hall B, Catteau J, Lloyd AR, et al. Differences in cervicovaginal fluid in women with preterm premature pupture of
in amniotic fluid and maternal serum cytokine levels in early midtrimester membranes: a comparison of amniotic fluid samples. Reprod Sci. 2016;
women without evidence of infection. Cytokine. 2008;44:78–84. 24:142–7.
15. Calhoun DA, Chegini N, Polliotti BM, Gersting JA, Miller RK, Christensen RD. 37. Park KH, Kim SN, Oh KJ, Lee SY, Jeong EH, Ryu A. Noninvasive prediction of
Granulocyte colony-stimulating factor in preterm and term pregnancy, intra-amniotic infection and/or inflammation in preterm premature rupture
parturition, and intra-amniotic infection. Obstet Gynecol. 2001;97:229–34. of membranes. Reprod Sci. 2012;19:658–65.
16. Cobo T, Jacobsson B, Kacerovsky M, Hougaard DM, Skogstrand K, Gratacos
E, et al. Systemic and local inflammatory response in women with preterm
prelabor rupture of membranes. PLoS One. 2014;9:e85277.
17. Skrablin S, Lovric H, Banovic V, Kralik S, Dijakovic A, Kalafatic D. Maternal
plasma interleukin-6, interleukin-1beta and C-reactive protein as indicators
of tocolysis failure and neonatal outcome after preterm delivery. J Matern
Fetal Neonatal Med. 2007;20:335–41.